go-v.orgdoi

ESTIMATING THE INTERVAL PRECEDING DIAGNOSIS OF TYPE 2 DIABETES FROM THE TIME-COURSE OF RETINOPATHY PREVALENCE

Massimo PORTA, Paola Dalmasso, Giulia Curletto, E. Coppo, A. Raviolo, Elena Sitia, S. Albani · Institutional Research Information System University of Turin (University of Turin) · 2013

  • DOI 10.5301/ejo
  • Publisher University of Turin
  • Type conference-paper
  • Issued 2013

Not cached yet. This record was fetched live for this request and is stored after the lookup. Reopening the page should serve the cached copy.

Institutional Research Information System University of Turin (University of Turin) · University of Turin · conference-paper · 2013

Abstract

Introduction. Type 2 diabetes (T2DM) has a slow onset and remains undiagnosed for years. This interval is estimated longer than 10 years, by extrapolating the correlation between known diabetes duration and prevalence of diabetic retinopathy (DR). However, those calculations were based upon weak assumptions: 1) that such correlation is linear, 2) any degree of DR was considered, including mild lesions observed in 10% of non diabetic individuals, 3) uncertain definition of T2DM, 4) time from onset of diabetes to appearance of DR arbitrarily set at 5 years. Purpose. To estimate the interval preceding diagnosis of T2DM after correcting the above assumptions Methods.

12,074 patients (35,545 screenings) were stratified into Younger-Onset (YO) if age at onset was <30 or Older-Onset (OO, ≥30) and currently on insulin treatment (IT) or not (NIT). DR was classified as mild or more severe (AnyDR) and moderate or more severe (Mod-DR). The best-fits between known duration of diabetes and prevalence of DR were calculated by the coefficient of determination (R2) and Akaike’s information criterion (AIC).

Time preceding diagnosis of T2DM was estimated by adding the negative intercept on the horizontal axis of the best-fitting correlation line between known duration of OO-NIT and prevalence of DR to the positive intercept of the bestfitting correlation between duration of YO-IT and appearance of DR. Results. There were 7,298 OO-NIT (equivalent to T2DM), 1,719 with AnyDR and 685 with ModDR, and 1,725 YO-IT (equivalent to T1DM), 756 with AnyDR and 385 with ModDR. In the OO-NIT the best-fits were a quadratic model suggesting appearance of AnyDR 3.89 years before diagnosis of T2DM and a linear model with ModDR appearing 2.66 years before diagnosis.

The onset of AnyDR following onset of YO-IT could not be timed, whereas a quadratic model suggested appearance of ModDR 3.29 years after its onset. The resulting estimate was (2.66+3.29) 5.95 years preceding T2DM diagnosis. Conclusions. Use of appropriate fitting models and stratification by treatment and DR severity lowers the estimate of unknown duration before diagnosis of T2DM to about 6 years, which is more realistic than previous estimates.

Open the resolved locator · 10.5301/ejo

Metadata licence: OpenAlex metadata (CC0)

Registry
OpenAlex DOI index + API
Fetched
2026-11-20 02:15 UTC
Cache state
not cached yet (bulk-index+openalex)
Locator
http://hdl.handle.net/2318/139230
Metadata licence
OpenAlex metadata (CC0)
Scope
Metadata as returned by the resolver; abstracts are never generated here.